跳到主要内容

聊天补全

POST /v1/chat/completions

核心对话接口，兼容 OpenAI Chat Completions 格式，支持多模态输入、流式输出、工具调用与推理模型。

请求​

curl https://test.kailin.com.cn/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-你的令牌" \
  -d '{
    "model": "deepseek-v4-1-flash-261017",
    "messages": [
      { "role": "system", "content": "你是一个简洁的助手" },
      { "role": "user", "content": "你好" }
    ],
    "stream": false
  }'
请求REQUEST
POSThttps://test.kailin.com.cn/v1/chat/completions
响应RESPONSE
填写参数并发送请求，响应会显示在这里

请求体参数​

参数类型必填说明
modelstring是模型名称，见 列出模型
messagesarray是对话消息数组，见下方消息结构
streamboolean否是否流式输出（SSE），默认 false
temperaturenumber否采样温度 0–2，默认 1，越低越确定
top_pnumber否核采样，与 temperature 二选一调整
max_tokensnumber否最大生成 token 数（部分新模型改用 max_completion_tokens）
stopstring / array否停止序列，最多 4 个
toolsarray否工具（函数）定义列表，见工具调用
tool_choicestring / object否auto / none / required / 指定工具
response_formatobject否{"type": "json_object"} 或 {"type": "json_schema", ...}
seedinteger否随机种子，尽量复现输出
frequency_penaltynumber否-2.0–2.0，降低重复
presence_penaltynumber否-2.0–2.0，鼓励新话题
reasoning_effortstring否推理模型思考力度：low / medium / high
userstring否终端用户标识，便于风控

messages 消息结构​

字段类型说明
rolestringsystem / user / assistant / tool
contentstring / array字符串，或多模态内容数组（见多模态输入）
namestring可选，参与者名称
tool_callsarrayassistant 消息中模型发起的工具调用
tool_call_idstringtool 角色消息必填，对应被响应的调用 ID

响应（非流式）​

{
  "id": "chatcmpl-B3xK1abc123",
  "object": "chat.completion",
  "created": 1730000000,
  "model": "deepseek-v4-1-flash-261017",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "你好！有什么可以帮你的吗？"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 20,
    "completion_tokens": 12,
    "total_tokens": 32
  }
}
字段说明
choices[].message.content模型回复文本
choices[].finish_reasonstop 正常结束 / length 达到上限 / tool_calls 需要执行工具
usagetoken 计量，计费依据
usage.completion_tokens_details.reasoning_tokens推理模型返回的思考 token 数

流式响应​

"stream": true 时返回 SSE 事件流：

data: {"id":"chatcmpl-xxx","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}

data: {"id":"chatcmpl-xxx","choices":[{"index":0,"delta":{"content":"你好"},"finish_reason":null}]}

data: {"id":"chatcmpl-xxx","choices":[{"index":0,"delta":{},"finish_reason":"stop"}]}

data: [DONE]

增量内容在 choices[].delta.content，最后以 data: [DONE] 结束。最后一个 chunk 的 usage 字段可能携带 token 用量（取决于上游）。

多模态输入​

content 传数组即可混合文本与图片/音频。图片支持 URL 或 Base64 Data URL：

{
  "model": "gpt-4o-mini",
  "messages": [
    {
      "role": "user",
      "content": [
        { "type": "text", "text": "这张图里有什么？" },
        {
          "type": "image_url",
          "image_url": {
            "url": "https://upload.wikimedia.org/wikipedia/commons/3/3c/Shaki_waterfall.jpg"
          }
        }
      ]
    }
  ]
}
内容块类型字段说明
texttext文本
image_urlimage_url.url图片 URL 或 data:image/jpeg;base64,...
input_audioinput_audio.data / input_audio.formatBase64 音频（wav / mp3），需模型支持

工具调用​

  1. 请求中通过 tools 声明工具，模型返回 tool_calls；
  2. 客户端执行工具后，把结果以 role: "tool" 消息追加，再次请求。
{
  "model": "deepseek-v4-1-flash-261017",
  "messages": [{ "role": "user", "content": "上海今天天气怎么样？" }],
  "tools": [
    {
      "type": "function",
      "function": {
        "name": "get_weather",
        "description": "查询指定城市当前天气",
        "parameters": {
          "type": "object",
          "properties": {
            "city": { "type": "string", "description": "城市名" }
          },
          "required": ["city"]
        }
      }
    }
  ]
}

模型返回：

{
  "choices": [
    {
      "message": {
        "role": "assistant",
        "tool_calls": [
          {
            "id": "call_abc123",
            "type": "function",
            "function": {
              "name": "get_weather",
              "arguments": "{\"city\": \"上海\"}"
            }
          }
        ]
      },
      "finish_reason": "tool_calls"
    }
  ]
}

执行后把结果回传：

{
  "model": "deepseek-v4-1-flash-261017",
  "messages": [
    { "role": "user", "content": "上海今天天气怎么样？" },
    {
      "role": "assistant",
      "tool_calls": [
        {
          "id": "call_abc123",
          "type": "function",
          "function": { "name": "get_weather", "arguments": "{\"city\": \"上海\"}" }
        }
      ]
    },
    { "role": "tool", "tool_call_id": "call_abc123", "content": "晴，26°C" }
  ]
}

使用 SDK​

from openai import OpenAI

client = OpenAI(
    api_key="sk-你的令牌",
    base_url="https://test.kailin.com.cn/v1",
)

stream = client.chat.completions.create(
    model="deepseek-v4-1-flash-261017",
    messages=[{"role": "user", "content": "写一首关于秋天的短诗"}],
    stream=True,
)
for chunk in stream:
    delta = chunk.choices[0].delta
    if delta.content:
        print(delta.content, end="", flush=True)
import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: 'sk-你的令牌',
  baseURL: 'https://test.kailin.com.cn/v1',
});

const stream = await client.chat.completions.create({
  model: 'deepseek-v4-1-flash-261017',
  messages: [{ role: 'user', content: '写一首关于秋天的短诗' }],
  stream: true,
});
for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.delta?.content ?? '');
}
备注

推理类模型（如 o 系列、DeepSeek-R1）的思考过程可能在 reasoning_content 或 delta.reasoning_content 字段返回，具体取决于上游渠道。